powernow-k8 refuses to load

[Jason Toffaletti]

[4294692.044000] powernow-k8: Found 1 AMD Athlon 64 / Opteron processors (version 1.50.4)
[4294692.044000] powernow-k8: MP systems not supported by PSB BIOS structure

2.6.15-16-686 #1 SMP PREEMPT Mon Feb 20 17:26:04 UTC 2006 i686 GNU/Linux

was working fine in breezy. I tried pci=routeirq and it didn't help.

[Jason Toffaletti (jason)]

There is also an ubuntuforum thread about this bug:

http://www.ubuntuforums.org/showthread.php?p=776465

[Jason Toffaletti (jason)]

attaching dmesg

[Jason Toffaletti (jason)]

I installed linux-image-2.6.15-15-386_2.6.15-15.21_i386.deb from the flight-4 iso and it does not have this issue, so something relatively recent caused this problem.

[Idan Gazit (idan-fastmail)]

Same exact issue occurs with k7 kernel images (32-bit dapper flight 4 on amd64 machine):

uname -a:
Linux shiny 2.6.15-16-k7 #1 SMP PREEMPT Mon Feb 20 18:14:55 UTC 2006 i686 GNU/Linux

[Fabien Chaillou (fchaillo)]

Same for me ... i opened a bug '#32538) that i have now declared as a duplicate of this one
Hope it will be fixed soon

[lune001 (n-lussier)]

Same things append here, with a 2.6.15.17-k8 kernel

[Henning (henning-schroeder)]

I got the same problem with a mobile sempron 3100+ processor and kernel 2.6.15-17-k7 included in Dapper.

[Helmut Duregger (helmutduregger)]

Same here, also with kernel 2.6.15-17-k7 on Athlon 64 3200+.

[Helmut Duregger (helmutduregger)]

still in 2.6.15-17.26 of package 2.6.15-17-k7. I know it hasnt been asigned yet :-D

[rubinstein (rubinstein)]

And the same with 2.6.15-18-k7. Changing also Status to confirmed since I think there are enough people who get the same behaviour.

[rubinstein (rubinstein)]

Confirmed by lots of people.

[Fabien Chaillou (fchaillo)]

I think it is not only for the k7 version of the kernel but for every version ... i tried with the amd64-generic and amd64-k8 and i still got the same problem even with the latest 2.6.15-18

Hope it will be fixed soon :)

[sorchuchu (riccardo-leccese)]

 have the same exact problem on a machine with Dapper on AMD64 Turion.

That's what dmesg | grep powernow says about:
[ 31.117297] powernow-k8: Found 1 AMD Athlon 64 / Opteron processors (version 1.50.4)
[ 31.117307] powernow-k8: MP systems not supported by PSB BIOS structure

Before i was running an Ubuntu Breezy and everything was working fine.

I suppose this problem it's related to the kernel that enable the smp also on a machine that actually doesn't have double processor.

I really hope the problem will be fixed soon.

[Matthew (shadesfox)]

I've been having this issue too. I've done a bit of kernel hacking and discovered that in my case powernow-k8 refuses to load is that the function num_possible_cpus() returns 3. Why it wants to say my computer can have 3 processors is beyond me.

Out of curiosity I built a module with num_present_cpus() instead of num_possible_cpus() and it works shiny, though it has been pointed out to me that this might not be best.

http://launchpad.net/distros/ubuntu/+source/linux-source-2.6.15/+bug/33846
(that bug should probably be marked a duplicate)

It still isn't working for me with the updated kernel, but I'm rebuilding the kernel now minus CPU hotplug. Given the behavior it seems like a bug in there (or at least a bug uncovered by cpu hotplug).

[Fabien Chaillou (fchaillo)]

I just applied your patch and rebuild only the powernow-k8 kernel module and replaced the one from the main kernel with mine.

It works well except that message

[ 44.503304] powernow_k8: no version for "struct_module" found: kernel tainted which i think has nothing to do with the patch.

Thanx anyway i'll get my ears back now with the drop in noise in my room :)

[Matthew (shadesfox)]

Good to hear it worked for someone else too.

Though I think that the right solution would be to track down why num_possible_cpus() returns strange numbers (like 3).

Though I have found that turning off cpu hotplug in the kernel config will solve the problem as well, so it is definitely related to cpu hotplug.

So the order of solution I would like to see is:

1. Track down what is wrong with num_possible_cpus() (I'm trying but I don't think I know enough about the kernel to really be able too do that)

2. Disable cpu hotplug in the kernel. It is marked at alpha on the source forge page and it looks like it has it's hands pretty deep in the kernel. With this in mind it seems like it would be a bad idea to enable cpu hotplug by default.

3. Apply the patch I have in the other bug report. I would probably cause bad things to happen if you tried to hotplug a cpu while powernow is running, but it seems to be a reasonable stop gap measure. Then again, I'm not a real kernel dev.

[Bruce Cowan (bruce89-deactivatedaccount)]

Mine's alright :
[ 54.777625] powernow-k8: Found 1 AMD Athlon 64 / Opteron processors (version 1.50.4)
[ 54.777659] powernow-k8: 0 : fid 0xc (2000 MHz), vid 0x6 (1400 mV)
[ 54.777662] powernow-k8: 1 : fid 0xa (1800 MHz), vid 0x8 (1350 mV)
[ 54.777665] powernow-k8: 2 : fid 0x2 (1000 MHz), vid 0x12 (1100 mV)
Linux Droopy 2.6.15-19-amd64-k8 #1 SMP PREEMPT Mon Mar 20 17:03:51 UTC 2006 x86_64 GNU/Linux

[Jason Toffaletti (jason)]

still having this issue with: 2.6.15-19-k7

[4294959.425000] powernow-k8: Found 1 AMD Athlon 64 / Opteron processors (version 1.50.4)
[4294959.425000] powernow-k8: MP systems not supported by PSB BIOS structure

[Idan Gazit (idan-fastmail)]

Same here, still not fixed in 2.6.15-19-k7.

[Jason Toffaletti (jason)]

Fixed in 2.6.15-20-k7, thank you!

[jeremyh (jeremy-jeremyh)]

jeremy@ubuntu:~$ dmesg | grep powernow
[4294689.517000] powernow-k8: Found 1 AMD Athlon 64 / Opteron processors (version 1.50.4)
[4294689.521000] powernow-k8: BIOS error - no PSB or ACPI _PSS objects

Linux ubuntu 2.6.15-20-k7 #1 SMP PREEMPT Tue Apr 4 19:26:13 UTC 2006 i686 GNU/Linux

AMD Athlon64 3000+

[Henning (henning-schroeder)]

Unfortunately it is not fixed for me with linux-image-2.6.15-20-k7:

[4294705.557000] powernow-k8: Found 1 AMD Athlon 64 / Opteron processors (version 1.50.4)
[4294705.557000] ACPI-0517: *** Error: Method parse/execution failed [\_PR_.CPU0._PSS] (Node ddfd3c20), AE_AML_PACKAGE_LIMIT
[4294705.563000] powernow-k8: BIOS error: numpst must be 1
[4294705.587000] ACPI-0517: *** Error: Method parse/execution failed [\_PR_.CPU0._PSS] (Node ddfd3c20), AE_AML_PACKAGE_LIMIT
[

[Ben Collins (ben-collins)]

Looks like a totally different bug. Your issue seems to be a BIOS bug. Something I cannot fix.

[Helmut Duregger (helmutduregger)]

thanks, it works now!!!

powernow-k8: Found 1 AMD Athlon 64 / Opteron processors (version 1.50.4)
powernow-k8: 0 : fid 0x2 (1000 MHz), vid 0x12 (1100 mV)
powernow-k8: 1 : fid 0xa (1800 MHz), vid 0xa (1300 mV)
powernow-k8: 2 : fid 0xc (2000 MHz), vid 0x6 (1400 mV)
powernow-k8: 3 : fid 0xe (2200 MHz), vid 0x2 (1500 mV)
cpu_init done, current fid 0xe, vid 0x2

 Linux version 2.6.15-21-k7 (buildd@rothera) (gcc version 4.0.3 (Ubuntu 4.0.3-1ubuntu5)) #1 SMP PREEMPT Fri Apr 21 17:10:51 UTC 2006

[Steven (etienne-pallier)]

I have the same problem with a bi-pro AMD 64 running fedora 5 :

$ dmesg |grep powernow
powernow-k8: Found 2 AMD Athlon 64 / Opteron processors (version 1.60.2)
powernow-k8: MP systems not supported by PSB BIOS structure
powernow-k8: MP systems not supported by PSB BIOS structure

I updated from kernel 2.6.15 to kernel 2.6.17 (2.6.17-1.2157_FC5 #1 SMP)

but the problem is still there...

Anyway, the server seems to be working fine. Should I worry ?
Should I rebuild the source kernel without the option "cpu hotplug" ?

[Jason Toffaletti (jason)]

I think this bug can be closed now.

[Jason Toffaletti (jason)]

closing.

[mabovo (mabovo)]

Sorry to post here but I got this error on Jaunty with 2.6.26-11.

[mabovo (mabovo)]

2.6.28-11.

[Haggai Eran (haggai-eran)]

I'm also affected by this bug, in Hardy, with kernel 2.6.24-24-generic (64bit).